Victoria Mercy Kataike1, Patricia M Desmond2,3, Christopher Steward1,3, Bruce CV Campbell4,5, Felix Ng4,5, and Vijay Venkatraman1,3
1Radiology, University of Melbourne, Melbourne, Australia, 2University of Melbourne, Melbourne, Australia, 3Medical Imaging, The Royal Melbourne Hospital, Melbourne, Australia, 4Neurology, The Royal Melbourne Hospital, Melbourne, Australia, 5Medicine, The Royal Melbourne Hospital, Melbourne, Australia
Synopsis
Keywords: Stroke, Oxygenation, Oxygen Extraction fraction, DWI reversal, infarct reversal, ischemic stroke
Motivation: In patients with ischemic stroke, ischemic tissue reversal after reperfusion therapy has been studied structurally, but metabolic changes have not been widely researched.
Goal(s): Assess the ability of a recently developed algorithm to quantify oxygen extraction fraction in ischemic tissue that reverses and what does not reverse.
Approach: Using multi-echo gradient images, oxygen extraction fraction and magnetic susceptibility maps were generated for 26 ischemic stroke patients. Values were extracted respectively at 24 to 72 hours in the ischemic tissue that reversed and tissue that did not.
Results: Significant differences in oxygen extraction were observed in tissue that reversed versus tissue that did not.
Impact: Differentiating oxygen extraction fraction
values in ischemic tissue at an early stage of ischemic stroke can potentially
inform neuroprotection strategies for clinicians to reduce post-reperfusion
secondary injury.
Introduction
The phenomenon of diffusion weighted imaging (DWI) reversal has been reported in
many studies in successfully reperfused ischemic stroke patients when comparing pre-treatment DWI and post-treatment DWI1. On the other hand, tissue that remained abnormal on DWI
after reperfusion but subsequently normalised may represent successful salvage
from secondary injury. Although previous studies on ischemic tissue with DWI
reversal have focused on its frequency, structure, and predictors, little
research has been conducted on the associated metabolic changes. Oxygen
Extraction Fraction (OEF) is considered to be a measure of
hemodynamic function and a physiological parameter of the brain's energy metabolism2. Quantitative Susceptibility Mapping (QSM) is regarded as a measure of iron metabolism3. Several MR-based quantitative
techniques have been recently proposed to estimate OEF using dual-calibrated
fMRI and multi-echo Gradient echo (mGRE) images. In this study, we used mGRE images to generate
a Quantitative Susceptibility Mapping (QSM) map for iron metabolism. For OEF mapping, we used an estimation technique (QSM+qBOLD-based
mapping with temporal clustering, tissue composition, and total variation – QQ+CCTV) that integrates phase and magnitude information from mGRE images4. In this study, we sought to observe
and assess the utility of the QQ+CCTV mapping technique in characterising the
differences between tissue that exhibited reversal during the post-treatment
period (i.e. abnormal on DWI after treatment but eventually normalised), and
tissue that did not reverse in patients who underwent successful endovascular
thrombectomy.Methods
Twenty-six
ischemic stroke patients (12 female, mean age 63.8 ± 14.2 years) underwent endovascular thrombectomy (EVT) and
thereafter magnetic resonance imaging (MRI) scans at 24 to 72 hours (tp1), and
3 months (tp2) after treatment to obtain DWI, fluid-attenuated inversion
recovery (FLAIR), and multi-echo gradient echo (GRE) images. The median time
from EVT to the tp1 scan was 46.77 (interquartile range: 27.70 – 71.48) hours. OEF
maps and QSM maps were generated using the QQ+CCTV and morphology-enabled
dipole inversion (MEDI) pipelines, respectively4. Infarct regions were manually delineated on DWI images at
tp1 and FLAIR images at tp2 by a neuroradiologist. Additionally, mirror ROIs were
automatically generated and manually verified. Regions of cerebrospinal fluid on
tp2 FLAIR were excluded by segmentation to adjust for post-stroke atrophy as
per previous published methods. ROIs delineated in
tp2 images were co-registered to tp1 images using the Advanced Normalization
Tools (ANTs)5. Using the FSL
library tools, we identified regions of Infarct Reversal (defined as regions
with DWI restriction at 24 to 72 hours [tp1] but normal appearance on 3 month
FLAIR [tp2]) and Persistent Infarct (defined as regions with DWI restriction at
tp1 and hyperintensity on tp2 FLAIR. The Infarct Reversal and Persistent
Infarct ROIs were then mapped onto the QSM and OEF maps and values were
extracted at tp1. Statistical comparison between the ROIs was conducted using
paired t-tests, with a significance threshold set at 0.05. Results
The OEF at tp1 of
the Infarct Reversal region was significantly higher than the OEF of the Persistent
Infarct (23.71 ± 6.70% vs 21.65 ± 6.94%, p < 0.001).
The QSM of the Infarct Reversal region was similar to that of the Persistent
Infarct (-1.16 ± 14.63 ppb vs -2.50 ± 16.59 ppb, p = 0.54).
Both the OEF of the Infarct
Reversal region and that of the Persistent Infarct were significantly lower
than their corresponding control mirror homolog ROI (28.19 ± 7.10%) (p < 0.001 for both). The QSM in both Infarct Reversal
region and the Persistent Infarct were similar to the control (0.26 ± 15.05 ppb) (p = 0.54 and p = 0.28 respectively).Discussion
The QQ+CCTV
technique characterized differences in oxygen extraction in tissue with DWI
abnormality during the acute period that subsequently proceeded to long-term reversal
versus tissue that did not. OEF values in the reversed region were less
deranged at 24 to 72 hours than those of the tissue that evolved to persistent
infarction, consistent with metabolically compromised but viable tissue despite
exhibiting DWI restriction. We demonstrate the potential for OEF imaging of the
acute infarct to identify salvageable but radiologically abnormal tissue that
may become a therapeutic target of
neuroprotection strategies to prevent post-reperfusion secondary injury. Not
surprisingly, there were no changes in iron levels between tissue that reversed
region versus the persistent infarct during the acute period, as detectable
changes associated with neurodegeneration is expected to occur beyond the 24 to 72 hour time point.Conclusion
The QQ+CCTV technique
provides unique insight into the acute metabolic profile of ischemic tissue and
underlying pathophysiology of DWI reversal in the post-reperfusion period. OEF
evaluation of the ischemic tissue shortly after treatment may predict long-term
tissue fate.Acknowledgements
Australian Brain Foundation, Australian and New Zealand Association of
Neurologists, Austin Medical Research Foundation, Royal Australasian College of
Physicians, and the Royal Melbourne Hospital Neuroscience Foundation.References
1. Nandakumar N, John RF, Steven W, Jośe GM. Reversible diffusion-weighted imaging lesions in acute ischemic stroke. Neurology. 2020;94(13):571. doi:10.1212/WNL.00000000000091732.
2. Jian Z, Wang X, Tian M, et al. Review of the Research Progress of Human Brain Oxygen Extraction Fraction by Magnetic Resonance Imaging. Oxid Med Cell Longev. 2022;2022:4554271. doi:10.1155/2022/45542713.
3. Langkammer C, Schweser F, Krebs N, et al. Quantitative susceptibility mapping (QSM) as a means to measure brain iron? A post mortem validation study. NeuroImage. 2012-09-01 2012;62(3):1593-1599. doi:10.1016/j.neuroimage.2012.05.0494.
4. Cho J, Spincemaille P, Nguyen TD, Gupta A, Wang Y. Temporal clustering, tissue composition, and total variation for mapping oxygen extraction fraction using QSM and quantitative BOLD. Magn Reson Med. Nov 2021;86(5):2635-2646. doi:10.1002/mrm.288755.
5. Avants B, Tustison N, Song G. Advanced normalization tools (ANTS). Insight J. 11/30 2008;1–35doi:10.54294/uvnhin